ai-ml-lifecycle
Route AI evaluation, ML training and inference engineering, and model operational readiness; use security review for authority-boundary threats.
SKILL.mdSKILLS/ai-ml-lifecycle/SKILL.md
GitHubAI And ML Lifecycle
Choose one primary mode and read only its workflow. Add another mode only when the requested evidence crosses lifecycle stages.
| Mode | Use when | Workflow |
|---|---|---|
ai-evals |
Design cases, datasets, rubrics, judges, thresholds, or behavioral regression evidence. | AI evaluations |
ml-engineering |
Design or review features, training, evaluation, packaging, inference, or integration behavior. | ML engineering |
mlops-readiness |
Review registry, promotion, reproducibility, deployment, monitoring, drift, or rollback controls. | MLOps readiness |
Do not load sibling workflows or artifacts prospectively. Keep claims tied to datasets and executed evidence; keep production promotion and rollback behind named owner approval.